Abstract: Subcritical reactivity measurement technology can effectively prevent accidental criticality of nuclear reactors and shorten the time of physical start-up tests, thus providing significant advantages in both safety and economic efficiency. For subcritical systems with an external neutron source, the steady-state neutron flux distribution consists of a superposition of the fundamental mode and higher-order harmonics. This harmonic superposition effect becomes particularly prominent under deep subcritical conditions, which severely restricts the accuracy of reactivity values obtained by the traditional neutron source multiplication method (NSM) and the modified neutron source multiplication method (MNSM). To overcome this drawback, a fundamental mode extraction neutron source multiplication method (FENSM) was derived in this paper, based on the physical interpretation that reactivity describes the neutron multiplication characteristics of the neutron flux in the fundamental mode state. Simulation experiments were carried out based on a uranium-water lattice zero power facility. A series of subcritical cases with different subcriticality levels were established, and comparative studies were conducted to analyze the variation trends of reactivity accuracy obtained by NSM, MNSM and FENSM as the system becomes deeper subcritical. The results show that the proportion of the fundamental mode neutron flux component decreases with deepening subcriticality, and the fundamental mode fraction of the thermal group is higher than that of the fast group. And the deviations between the subcritical reactivity calculated by the three methods and the reference value all increase with the deepening of subcriticality. The accuracy of the three methods ranks in descending order as FENSM, MNSM and NSM. When the effective multiplication factor k_eff ranges from 0.833 63 to 0.978 16, the deviation between the subcritical reactivity obtained by FENSM and the reference value is less than 400 pcm. FENSM maintains satisfactory and stable accuracy over a wide range of subcriticality and is insensitive to the selection of the reference subcritical state. In addition, various subcriticality cases were achieved by modifying the fuel arrangement in the core region, and the applicability of the FENSM was further investigated under distorted global neutron flux profiles. By analyzing the numerical variation trends of each correction factor in the core center and reflector region, this study verifies that all correction terms are essential for distorted global neutron flux distributions. Furthermore, the FENSM still maintains high accuracy even though the fundamental neutron flux distribution differ significantly under different subcritical conditions. The research results of this paper can provide sufficient theoretical support and reliable reference for the follow-up experimental applications of the FENSM in uranium-water lattice zero power facilities.